Cognitive and Usability Engineering Methods for the Evaluation of Clinical Information Systems

View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Elsevier - Publisher Connector Journal of Biomedical Informatics 37 (2004) 56–76 www.elsevier.com/locate/yjbin Methodological Review Cognitive and usability engineering methods for the evaluation of clinical information systems Andre W. Kushniruka,* and Vimla L. Patelb a Information Technology Division, Department of Mathematics and Statistics, York University, Toronto, Canada b Laboratory of Decision Making and Cognition, Department of Biomedical Informatics, Columbia University, New York, NY Received 29 September 2003 Abstract Increasingly healthcare policy and decision makers are demanding evidence to justify investments in health information systems. This demand requires an adequate evaluation of these systems. A wide variety of approaches and methodologies have been applied in assessing the impact of information systems in health care, ranging from controlled clinical trials to use of questionnaires and interviews with users. In this paper we describe methodological approaches which we have applied and refined for the past 10 years for the evaluation of health information systems. The approaches are strongly rooted in theories and methods from cognitive science and the emerging field of usability engineering. The focus is on assessing human computer interaction and in particular, the usability of computer systems in both laboratory and naturalistic settings. The methods described can be a part of the formative evaluation of systems during their iterative development, and can also complement more traditional assessment methods used in summative system evaluation of completed systems. The paper provides a review of the general area of systems evaluation with the motivation and rationale for methodological approaches underlying usability engineering and cognitive task analysis as applied to health information systems. This is followed by a detailed description of the methods we have applied in a variety of settings in conducting usability testing and usability inspection of systems such as computer-based patient records. Emerging trends in the evaluation of complex information systems are discussed. Ó 2004 Elsevier Inc. All rights reserved. 1. Introduction deployment of information systems as well as ensuring that the process of design of health care systems leads to Effective evaluations of health care information sys- effective systems [2]. In the general software industry it is tems are necessary in order to ensure that systems ade- increasingly recognized that continued evaluation is quately meet the requirements and information needed throughout the system development lifecycle, processing needs of users and health care organizations. from early design to summative testing, in order to en- A range of approaches have been used in the assessment sure final products meet expectations of designers, users, of information systems with the initial focus on sum- and organizations [2,3]. A variety of cognitive ap- mative evaluation with the objective of assessing how proaches to the assessment of health information sys- well-completed systems meet a set of pre-defined goals tems have been developed based on ideas from cognitive regarding issues of functionality, safety, and impact on and usability engineering. The methods typically borrow outcome measures such as cost of health care and work from an interdisciplinary perspective and draw from a efficiency [1]. However, in recent years an additional number of areas including cognitive psychology, com- focus has emerged: the development of approaches to puter science, systems engineering, and the field of us- evaluation that can be used in the iterative evaluation of ability engineering. Usability can be broadly defined as systems during their development (i.e., formative evalu- the capacity of a system to allow users to carry out their ation), with the objective of improving the design and tasks safely, effectively, efficiently, and enjoyably [4,5]. In recent years the field of usability engineering has * Corresponding author. Fax: +1-416-736-5392. emerged to address the need for the application of E-mail address: [email protected] (A.W. Kushniruk). scientific methods for improving system development 1532-0464/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.jbi.2004.01.003 A.W. Kushniruk, V.L. Patel / Journal of Biomedical Informatics 37 (2004) 56–76 57 and in particular human–computer interaction [6,7]. applied in real work tasks [11,12]. Along these lines, The profound influence of information systems on Rasmussen, Pejtersen, and Goodstein [11] argue that: cognitive processes of the users is documented in the ‘‘The fast change of pace in available technology makes it diffi- literature [8]. However, conventional summative and cult to develop appropriate information systems through an in- outcome-based evaluations are lacking in their ability to cremental updating of existing concepts. Instead design (and describe these potentially important effects of computer redesign) has to be based on a conceptual framework capable systems on human cognitive processes. In this paper we of supporting the analysis of work systems, and the prediction will focus on methods of evaluation emerging from of the impact of new information systems, as well as the evaluation of the ultimate user–system interaction. By necessity, such cognitive and usability engineering that can be applied a framework has to integrate modeling concepts from many dif- during a systemÕs development to provide feedback and ferent disciplines, such as engineering, psychology, and the cog- guidance for its ultimate design [9]. The generic meth- nitive, management, information and computer sciences.’’ odological approach has also been used to assess the impact of implemented systems on human reasoning In general, the methods represent a shift from a focus and decision making processes. In addition, variations on the design of software and system components to of the methods described can also be used for assessing gaining a better understanding of the interaction be- the information and processing needs of users of health tween health care information systems and end users in information systems [8]. conducting day-to-day tasks. We begin the next section Cognitive and usability engineering approaches to the of the paper with a discussion of some limitations of assessment of health information systems involve: (a) conventional approaches to evaluation in order to situ- characterizing how easily a user can carry out a task using ate our work in the application of complementary as- the system, (b) assessing how users attain mastery in us- sessment methods emerging from cognitive science and ing the system, (c) assessing the effects of systems on work usability engineering. practices, and (d) identifying problems users have in in- teracting with systems. Evaluation in this context in- 1.1. Need for new evaluation methodologies for health volves gathering information about the actual process of systems using a system by representative users performing representative tasks. The results of such evaluation can be Conventional outcome-based evaluations include used to improve features of the system prior to comple- quantitative assessments of the economic impact, accu- tion of the design of the system, or alternatively to assess racy, safety, and reliability of completed information the impact of fully implemented systems. Thus some of systems. In such studies comparisons are usually made the methods described in this paper blur the boundary between experimental group of subjects using a tech- between formative and summative forms of evaluation. nology (e.g., a clinical information system) and a control Along these lines, it has been argued that input from the group. Typically, such studies have pre-defined outcome behavioral, cognitive, and social sciences is essential for measures which are measured after the system has been not just critiquing completed systems, but also to provide deployed in some setting [1,13]. This may provide us essential input into the design process itself [5]. From this with information about aspects of system we are already perspective, the processes of design and evaluation can be aware of and interested in (e.g., effects of use of a system considered to be highly inter-related. In particular, dur- on hospital costs). If the outcome is positive or as ex- ing the iterative development of systems, evaluation pected, then our assumptions about the trial-based during design is essential in order to ensure that a new studies are correct. This provides valuable information information technology takes into account the needs and on the evaluation. However, if the outcome is negative, limitations of its end users rather than the preconceptions then there is often no way of knowing the reason for this of the designers with regard to user requirements. The outcome, using these methods of data collection. In user-centered approach to evaluation focuses on char- addition, many effects of health information technology acterization of cognitive skills involved in using a system can be considered to be ‘‘emergent’’—i.e., they are to carry out representative tasks and description of identified or discovered only through the monitoring of problems of users with varying levels of expertise and the process of system use [8,14]. Thus, ultimately tech- experience, as they learn how to use and master the sys- nology

Load more